Osmotic and Ionic Changes in Blood and Muscle of Migrating Salmonids

Osmoregulation in teleost fishes, whether they live in fresh water or in the sea, is a physiological activity very closely related to their survival. Yet in spite of the importance of osmoregulation surprisingly little is known about how fish deal with the physiological problem inherent in living in hypoosmotic and hyperosmotic environments. The ability of some fishes to regulate in both environments during migrations is of even greater interest. In his classical review of osmoregulation in aquatic animals, Krogh (1939) emphasizes the lack of direct evidence in the study of migrating diadromous fishes. He says: ‘...A small number of teleosts are able to stand a fairly rapid transference from fresh water to sea water and vice versa, and some of these undertake regular migrations between the two media, but the peculiarities which are responsible for this power are not at all clear in spite of the considerable amount of work spent upon the problem.’ Some advances can be recorded; thus Pyefinch (1955) in his review of the literature on the Atlantic salmon writes, ‘… the physiological and biochemical studies over the last twenty years or so have shown the advances that can be made now that general knowledge of endocrine organs and hormones has made it possible to investigate their function in fishes. Studies of the internal environment will not...provide the complete answer to the many outstanding problems of migration and smolt development, but further advances in this field, supplemented by appropriate investigation in the external environment, could provide a much more complete picture of these phases in the life-history of the salmon.’

In the years between the publication of these two accounts many data have been accumulated in the field of the physiology of migrating fishes, but relatively little information is available about the chemistry of the internal environment, about permeability, or about renal physiology, all of which would seem to be particularly relevant to a study of migration in diadromous fishes. The present study has been undertaken to provide information about the level of some ions in blood and muscle during the life-history and migratory periods of the salmon.

Such a migratory fish as the salmon might seem to be an ideal example for a study of euryhaline behaviour. In many ways this is so, but there are disadvantages. There is a long period in the life of the Atlantic salmon which is almost unknown, and when it is not possible to obtain any physiological material, namely the sea-water period. The adult fish almost always have to be studied in field conditions since they are large for the resources of most laboratory aquarium facilities, and for the same reason difficult to move from one place to another. Also, since they are economically important, the young stages are valuable as potential adults, and the adult ones valuable in themselves, so that the numbers required for physiological work are not always easy to obtain.

Fish of the species Salmo salar (L.) were used for most of the analyses; wild fish were collected from many different places in Great Britain, and some young fish were reared from eggs in laboratory conditions. Comparative analyses were made also on other Salmo species in Britain, i.e. S. trutta (L.) as freshwater ‘brown trout’ or marine ‘sea trout’ and 5. gairdnerii (Richardson), the rainbow trout (hatchery reared). Blood samples were obtained by cardiac puncture, with or without anaesthesia, without harming the fish. The most convenient anaesthetic was found to be ‘Metacaine’^* in a concentration of 1 : 10,000. Some other anaesthetics, viz. metycaine, trichlorethylene, carbon dioxide and urethane were tried but found to be less successful. Pre-spawning adult fish were netted at the mouths of rivers and blood was obtained from these by cardiac puncture immediately after the fish had been stunned.

Blood samples from very small fish were collected in soda glass cannulae, and immediately transferred to liquid paraffin in polythene sample tubes in the field. In the laboratory the blood samples were transferred to ‘Damarda’^†-lined watch-glasses which induced serum and a clot to separate. For larger fish, glass or nylon hypodermic syringes were used, and the blood was stored for short periods under paraffin in polythene bottles, and frozen with ‘Drikold’. As soon as possible the blood was centrifuged, and the plasma separated.

The freezing-point depression of the blood (Δ, °C.) was measured in very small quantities (o-ooi ml. or less) in silica capillaries, using Ramsay’s cryoscopic apparatus (Ramsay, 1949; Ramsay & Brown, 1955). Sodium, potassium and calcium were estimated by flame photometry (EEL photometer) using standard solutions and dilutions of a standard sea water for comparison and calibration. Carbonate was measured in freshly collected blood samples by the micro-diffusion method of Conway (Conway, 1957). Chloride was estimated either by a Conway micro-diffusion method, and/or by a potentiometric titration with silver nitrate.

Some attempts were made to estimate trimethylamine and its oxide, either after a steam distillation extraction or by a microdiffusion method. Although considerable amounts were found in muscle extracts, none could be detected in blood samples of either the freshwater or marine salmon.

This has been measured in blood from the young salmon embryos, in hatched alevins, fry, parr, smolts before and during their migration, in marine adults returning to spawn, in spawning adults and in the spawned kelts in both fresh water and sea water. The results are shown in Table 1 and in Fig. 1.

It can be seen that Δ_blood is at a constant level during the freshwater juvenile life, except for a rise during the first few weeks of life from 0·49° C. at hatching to 0·56° C. in the alevins and fry. This rise in blood concentration is apparent in the yolk-sac bearing alevin before feeding is initiated, and leads one to suppose that ionic regulation has begun already at this early stage. Busnel, Drilhon & Rafly (1946) measured the blood concentration in hatching eggs and found similar low figures for Δ_blood (0·49° C.) and a rise as the alevin develops (to 0·59° C. after 31 days). A rather higher figure (0·61° C.) is reported for ‘Vesicular alevins’ by Auvergnat & Sécondat (1941). The blood concentration rises slightly during the first year of freshwater life, but thereafter remains constant through the ‘parr-smolt’ transformation when the seaward migration begins.

When the parr metamorphoses into the smolt there is no significant change recorded for Δ_blood; that is, there is no increase in total blood concentration prior to entry into sea water, nor is there evidence for a ‘demineralization’ of the blood to stimulate a sea-going urge (cf. Fontaine & Callamand, 1940, 1948; Kubo, 1953). However, a greater variability in the results was found at this stage and this may represent a degree of sensitivity of the smolt (Fontaine & Callamand, 1948), or of hormone ‘un-balance’ in the fish at this time (Hoar, 1951). A similar variability (in sodium content of the blood) has been reported by Koch & Evans (1959) for wild migratory smolts.

If smolts were kept in fresh water beyond the time of migration, Δ_blood was found to be higher than in the migrating smolts. For example, fish from the River Piddle at Wareham (Dorset) kept in tanks of running hard water at Stevenage (Water Pollution Research Laboratory) for 12 months, and hatchery-reared fish of similar age from the Scottish Hydro-Electric Board’s hatchery at Invergarry, Inverness-shire, kept for a similar period in soft water, both showed this trend. There was no indication of a seasonal return to normal during the annual smolt migration; the rise observed could have been the result of ageing or of artificial feeding.

Very little is known of the life of the salmon in sea water and no material of this stage was obtained for the present study.

Adults from sea water, returning to the spawning grounds in the river, were captured in sea water at the mouth of the River Dee, Aberdeen. They had Δ_blood of the same order of that of other marine fishes, although it is about 10 % lower in concentration than that of many estimates reported in the literature (Black, 1957). This lower value is probably due to the difference in techniques for measuring Δ_blood. There is no reason to believe that the figures obtained from these fish did not represent the level of blood concentration in sea-living salmon. When the adult salmon reached the spawning ground and had been in a freshwater environment for 3–4 months, Δ_blood dropped to 0·61° C., which was a little higher than the level for juvenile salmon in fresh water. Spawned fish, although starving and exhausted, are still able to regulate since the blood does not fall below this level of concentration. However, ability to osmo-regulate is impaired to some degree, as the spawned kelts after 5-6 months of starving freshwater life cannot survive or maintain their normal blood concentration when transferred directly to sea water. Similar values are reported in the literature for adult salmon; Benditt, Morrison & Irving (1941) give Δ_bIood = 0·77° C. for salmon in I sea water (A = 0·87° C.), and figures reported for adult spawning salmon in fresh water are Δ_blood = 0·04–0·66° C. (Fontaine, 1954; Hoar, 1953).

Freezing-point measurements of the blood of two other Salmo species obtainable in Britain are shown in Table 2. No significant differences could be found between fish of the different species of similar age groups. The same general rise in blood concentration, observed in S. salar with increasing age, can be seen in S. trutta and S. gairdnerii.

Some figures for the Pacific genus Oncorhynchus from the literature seem to be considerably higher in range. Thus, Greene (1926) gives Δ_Wood = 0·61–0·67° C. for adults of O. tchawytscha in fresh water and 0·76° C. for these fish in sea water; and Kubo (1953), for O. masou gives 0·60–0·69° C. for fish in fresh water and 0·75–0·90° C. for fish in sea water. It is possible that this difference is a real one, related to the difference in species or environment, but it seems more probable that it reflects different techniques of freezing-point measurement. A summary of the figures reported in the literature for different species is given in Table 3.

The osmotic changes in the blood after transfer of the fish from fresh water to sea water is of particular interest in these migrating fish. The experimental transfer, to sea water and its dilutions, of young parr and smolts (1 and 2 years old) of three species of Salmo has been described in previous papers (Parry, 1958, 1960). In general, salmonids of age group 2+ years were able to tolerate a direct transfer from fresh water to full strength sea water. The changes in blood concentration following the transfer were measured by determining the freezing-point depression. It was found that for fish of the 2+ year age group 150–300 hr. were required for the fish to be able to regulate its blood concentration to the normal levels, except in the case of the migrating smolt of S. salar which regulated its blood within a 24 hr. period.

An attempt was made to find out what happened to smolts during similar conditions in a river. Smolts were trapped at a station (on the River Coquet, Northumberland) just above tidal water, and were transferred a mile downstream into tidal water. The blood concentration of these fish can be compared with that of fish living naturally in this environment, and also with fish from the same population transferred to running sea water in the laboratory. The temperature of the river water and of the running sea-water tanks was about 10° C. during these experiments (Table 4). Freezing-point measurements of the water in the tidal reach of the river gave values of Δ between 0·09 and 1·00° C. at different states of the tide. A few fish were caught (by rod and line) in the same part of the river and the blood was sampled; the experimental fish were held in a keep-box immersed in the river. It is interesting to note that the naturally occurring fish, which had been in the tidal reaches for an unknown time, had blood no more concentrated than that of fish in fresh water above the tide. Of the smolts confined in a keep-box, the ones sampled after three tides showed a higher blood concentration than those sampled after seven tides. Two explanations can be put forward for this: (1) that the abrupt transfer to a higher salinity could not be compensated by the osmoregulatory mechanisms; or (2) that diuretic loss of water consequent upon the confinement in the keep-box was preventing the reduction in urine flow which is a necessary part of the physiological adjustment of the fish to higher salinities.

Fish from the same population, transferred directly to running sea water, also showed a rise in blood concentration above the level normal for these fish in sea water, but this was controlled within 12 days.

A similar transfer of adult fish, but from fresh water to sea water, was attempted in January 1959 (Table 5). The design of this experiment was the same as the previous one, i.e. the fish were transported from the freshwater traps and confined in keep-boxes immersed in sea water. The sea water in Morecambe Bay where the experiment was made, had a freezing-point depression of Δ = 1·60° C. at the time of the experiment. The inability of the salmon kelts to osmoregulate is very much more marked than that of the salmon smolts in spite of a more favourable surface/volume ratio. Again, some of the increase in blood concentration could have been brought about by diuresis, but it is clear that whatever the cause, control had not been gained after 4 days in sea water. A contributory cause towards this inability to regulate could have been the low temperature of the water, which was at about 0° C. at this time, with ice forming on the inshore surface of the sea. A subsequent experiment in January 1960 with kelts kept in larger tanks of running sea water indicated a better degree of control with Δ_blood maintained at about 0·90–1·00° C., but in this case the sea water in More-cambe Bay was much diluted by rain water run-off, and the freezing-point of the sea water was only A = 1·20° C. The temperature in the 1960 experiment was similar to that of the earlier one.

A similar experiment with fresh-run adult fish moving into fresh water from sea water would be interesting, but so far is not available. The changes in the blood of a species of Pacific salmon, Oncorhynchus nerka, have been followed in natural conditions along the length of the Fraser River (Idler & Tsuyuki, 1958), and showed a gradual decline of blood concentration of about 5 % of the marine level.

Estimates of the total ions in the blood can be calculated from the measurement of blood from the relationship that 283 m-equiv. monovalent strong electrolyte per kg. water has Δ = 1° C. (Ramsay & Brown, 1955). The freezing-point measurements of blood discussed in the previous section thus show that the total ionic concentration rises after hatching from 148 m-equiv./kg. to about 160-170 m-equiv./kg. There is no rise in premigratory smolts, but there is a rise of about 12% to 200 m-equiv./kg. in smolts in tidal water or in sea water; this is maintained in the pre-spawning marine adult. The ionic concentration of the blood drops a little while the adult completes its spawning migration into fresh water.

Results of analyses are shown in Table 6 and summarized in Fig. 2. Sodium and chloride are numerically the most important ions of the total ionic concentration. In blood from juvenile fish in fresh water, sodium and chloride are present in sufficient concentrations to account for almost all the ions present. But in adult pre-spawning fish the chloride concentration is considerably less than the value expected from measurements of the freezing-point. This anionic deficiency is not covered by high carbonate. Analyses of carbonate show very similar levels in both sea-water and fresh-water fish (within the limits of experimental error, which are quite serious with regard to this ion). It might be thought that sea-water levels of carbonate would rise with low plasma chloride, or drop with high plasma chloride (Fontaine & Boucher-Firly’s hypothesis, 1934). The order of concentration of carbonate found, however, makes it unlikely that this ion can play much part in maintaining the total anionic concentration.

Both anionic and cationic organic constituents may be important, but analyses of plasma for trimethylamine oxide failed to indicate this substance in the plasma, though the methods employed should have been adequate (for the 1 ml. samples) if it had been present in significant quantities. No estimates of ammonia have been made. Further work is in progress to determine whether amino acids play an important part in osmoregulation during the transference from one salinity to another, as they do in some invertebrates (Shaw, 1958).

In general, the levels of individual ions follow the pattern expected from the freezing-point data. There is some indication that a size-concentration effect (Houston, 1959) is present; i.e. the concentrations of sodium and chloride in the plasma drop between that of the very young fingerlings and that of the 2-year smolt stage. Fresh-water and sea-water levels of plasma potassium seem to be about the same, and the range is well within that reported in the literature (Field, Elvehjem & Juday, 1943; Phillips & Brockway, 1958; Gordon, 1959a, b;Houston, 1959).

The plasma ion-levels of migrating smolts, and of smolts transferred artificially to sea-water, are interesting. Pre-migratory smolts in freshwater showed relatively variable figures for both freezing-point determinations and for inorganic analyses. This is shown by the larger standard errors of the figures in Tables 6 and 4. A similar variability is reported in plasma sodium levels of Atlantic salmon smolts (Koch & Evans, 1959). This high level of variability, however, is likely to be a reflexion of the curious physiological state of the smolt, rather than of direct osmotic significance.

The levels of inorganic ions in the muscle are of interest, from the osmotic point of view, in relation to their levels in the blood. In a migratory fish such as the salmon even greater interest hinges on the response of the muscle, as well as the blood, in conformity with the external salinity changes which the animal encounters during its migrations. There are two possible responses of muscle tissue to such changes:

(1) it can follow the osmotic and ionic changes in the blood, more or less closely, or

(2) it can behave differently and thus act as a reservoir of ions or water to buffer blood changes (Drilhon & Pora, 1936).

Analyses of the water content and some of the common inorganic ions which have been studied in blood plasma have been made for muscle (Table 7), and a pictorial summary is shown in Fig. 3.

If one compares, first of all, the muscle of freshwater spawning adults and the fresh-running sea-water adults, there are some quite marked differences to be seem.

The water content of the whole muscle has increased considerably, by nearly 15 %, during the passage of the fish upstream in a period of about 5 months. This increased water content of the muscle, however, does not mean that the whole has become more dilute or watery, as both the sodium and the potassium content of the whole muscle has increased, the sodium by 100%, and the potassium by 25%. This increase probably means that the extracellular space of the spawning fish in fresh water has increased. At the same time the calcium content, which might be expected to be inside the muscle fibres rather than in the extracellular fluid, remains the same. That these changes come after and not before entry into fresh water is shown by the analyses of muscle from a fish at the head of the tide on its way upstream. The analyses for this fish are substantially the same as those for the sea-water fish.

Analyses of muscle from migrating smolts in fresh water show an interesting contrast, in that although the water content of the muscle is high, the sodium is low. This could be interpreted in terms of a small extracellular space, but with the usual ionic content inside the muscle, because in this case the potassium (which will be mainly inside the cells) is high. A comparison of the two measurements of chloride in the muscle sub-stantiate this point. The figure for muscle from adult sea-living fish is about twice that of the smolt. Most of the chloride will be in the extracellular space, and thus the figures would represent a small extracellular space in the smolt, but a larger one in the marine adult.

Kelts which have been in fresh water for 9-10 months show different changes. In these fish the water content of the muscle is high (800‰) and the sodium content is mid-way between that of the young freshwater fish, and the adult spawning fish in fresh water. This could indicate an enlarged extracellular space (as in the spawning fish), but also a dilution of the muscle fibre contents, a conclusion borne out by the lower potassium content and by the drop in calcium.

Next we can consider the changes which follow the transfer of fish from one salinity to another when the fish is not in equilibrium osmotically. Two weeks after the transfer of migrating smolts to sea water the water content is scarcely reduced at all, nor does the sodium increase appreciably. Potassium, however, does drop. These changes could be interpreted as a movement of both water and ions out of the muscle fibres into a slightly increased extracellular space. It must be remembered that at the same time the blood becomes concentrated to some extent, so that the water movement could be an osmotic one and the ion extrusion from the fibres could be stimulated by the subsequent concentration within them; it is known that muscle fibres have well-developed powers of ion regulation. The slight increase in extracellular space could be the result of a time lag in these processes.

When the adult spent fish are subjected to the reverse change of salinity, the results are different. Kelts acclimatized to Morecambe Bay sea water (at a salinity equivalent to about two-thirds of oceanic water) showed a slight reduction in water content of the muscle, but retained the high sodium level. Potassium was lower than in spawning fish, but the same as in marine ones, while calcium was low, as in the spent fish. Thus, following our previous interpretation, the extracellular space is still enlarged, and this is indicated also by a high sodium content. The potassium is at the level characteristic of the fresh-run sea-living adult, so that in other respects the muscle is in equilibrium with the blood and with the external salinity, after this period of acclimatization. However, the figures for fish transferred to an artificial^* sea water for much shorter periods, show a pronounced disequilibrium 3 hr. after the transfer. The water content, perhaps surprisingly, had not changed very much from the 800 ‰ characteristic of the kelts in fresh water, so that although there was some osmotic depletion of water from the muscle into the blood, this is not so high as one might expect from the increase in blood concentration which follows such a transfer (Δ_blood changes from 0·63° C. to about 1·00° C.). The potassium does not change much in this period, but the sodium content does so very sharply. Thus while the volume of extracellular space only changes to a moderate degree, its ion concentration probably does increase, possibly by the movement of sodium from the blood plasma into the extracellular space. After a longer period in the artificial sea water, the water content of the muscle increases, while the sodium declines to a more normal level and the potassium increases a little. Presumably the extracellular space increases again at this time, as osmotic control in the blood is re-established by the movement of water into the extracellular space only; some of this water may come from the muscle fibres, since their potassium content is increased. That the fish is far from achieving equilibrium conditions is shown by the marked divergences of these analyses from those for fresh-running sea fish, as well as from a study of the blood changes at this time.

Values for the extracellular space are obviously of great importance in checking the assumptions made in this interpretation. Unfortunately, while such values can be obtained for small fish in aquarium conditions, e.g. in smolts being transferred from fresh water to sea water, it is very difficult to arrange an experiment to obtain values directly for the larger adult fish.

Similar increases in extracellular space of adult Salmo gairdnerii recently transferred from fresh water to sea water have been reported by Houston (1959) from similar data, but Gordon (1959a), studying salinity transfers in 5. trutta, thinks that the extracellular compartment is more or less constant. One of the most important differences between the experiments of these two authors and the ones described here is that the changes reported in this paper are those shown by the naturally migrating fish, at the relevant times of the year, so that the three sets of experiments may not be strictly comparable.

Sodium in the muscle of adult fish is reported to be subject to considerable seasonal and reproductive changes, e.g. in 5. trutta (Gordon, 1959a) and in Oncorhynchusnerka (Idler & Tsuyuki, 1958). Some of the change from 66 mM./kg. in spawning fish, to 45 mM./kg. in the post-spawned fish could be due to this. It is tempting to correlate the changes during the reproductive period with changes in availability of steroid hormones; Canadian workers have found high levels of cortisone and cortisol in adult fish just prior to spawning (17 times the normal human level), while in fish before entry into fresh water the level was only one-sixth of this (Idler, Ronald & Schmidt, 1959): Spalding (unpublished, reported in Chester Jones, 1956) showed that dosing brown trout with DCA or cortisone raised the sodium and lowered the potassium in muscle. Similar changes in the concentrations of steroids in S. salar have been found during the parr-smolt change (Fontaine & Hatey, 1954).

Trimethylamine oxide, while apparently absent from the blood in significant quantities, was found in the muscle to a variable extent. In post-migratory smolts still in fresh water, the concentration was 35 mg. %, in fresh-running marine adults it was 75 mg. %, and had risen as high as 250 mg. % in adult spawning fish in fresh water. Kelts in fresh water 3–4 months after spawning had very variable concentrations between 100 and 176 mg. %. The presence of any trimethylamine at all in the juvenile freshwater fish is interesting in view of the hypothesis that this substance is a characteristic of marine fish only, although it has previously been reported for freshwater fish (Anderson & Fellers, 1952). This trimethylamine is unlikely to be derived from the diet, but could represent the final stage of a basic metabolic pathway which is relatively inoperative in the juvenile fish. The measurements of trimethylamine in the adult fish are very interesting, in that the level is so high in the spawning fish. This must argue against an osmoregulatory function for this substance (Beatty, 1939) since one would expect the opposite change. It must be remembered that the spawning fish are starving, and perhaps the high level of trimethylamine present in the muscle arises from the metabolism of its own tissues during this period. The lower and extremely variable concentrations found in spent kelts could relate to their degree of debilitation.

Some specific points arising from this investigation may be mentioned. The anionic deficiency of the blood in the marine fish is of great interest osmotically, and attempts are being made to characterize this.

Nitrogen-containing compounds are possible contributors to the ion content of the plasma. Ammonia seems seldom to have been investigated in salmonids. Its concentration is likely to be low, however (reported as 0·104 mg. % in laked Salvelinus blood) since both the kidney and gill epithelia of teleosts are permeable to it. In teleosts generally the blood level is low. Trimethylamine or its oxide is a possible non-protein nitrogen constituent, but this was not found in any significant amounts in the plasma, even though it is present in the muscles. Betaine is another possibility which has not yet been investigated. Protein nitrogen, as amino acids, seems to be present in the blood in significant amounts. Hoar (1953) records ca. 70 mg./l. protein in ‘mature’ (?sea water) adults and a drop to 33 mg./l. in spawning fish. Recent analyses of amino acids in the blood of rainbow trout and salmon kelts indicate levels of amino acids between 30 and 60 mg./l.^* How this is related to the spawning migration and to the freshwater–sea-water transfer is not yet known.

Cholesterol is another organic constituent of blood which may be important osmotically, first, in a direct concentration effect, and secondly because of its water-binding affinities. In the Pacific salmon, Oncorhynchus nerka, the total cholesterol in the blood is reduced from nearly 600 mg. % at the beginning of the spawning migration to 200 mg. % at the point of spawning. The pre-spawning level is singularly high in comparison with warm-blooded animals (Idler & Tsuyuki, 1958). Similar high levels have been reported for other fish (carp, 662 mg. %, Field et al. 1943) and for other salmonids (in a Japanese river, probably O. keta, 150 mg. %, Okamura 1935). The relationship between cholesterol levels and water balance would be an interesting problem.

The possible presence of other, unspecified constituents in the plasma should not be overlooked. Recent investigations of the osmotic behaviour of some arctic fish have shown that these teleosts can raise the total osmotic concentration of the blood to equivalence with sea water. The extra blood concentration was not contributed by the common inorganic ions, nor by glucose, glycerol, proteins, urea or ammonia (Scholander, Van Dam, Kanwisher, Hammel & Gordon, 1957).

Some authors, notably Kubo (1953), Pickford & Atz (1957), Fontaine & Callamand (1948) and Fontaine (1951) have reported falling plasma and tissue concentrations of chloride in pre-migratory juvenile and adult fish, and have suggested that this is a characteristic of anadromous species. Fontaine & Callamand further suggested that the electrolyte depletion plays a role in the initiation of the seaward migration by producing increased activity in the fish. Investigations by Fontaine & Baraduc (1954), Fontaine, Lachiver, Leloup & Olivereau (1948), Fontaine & Leloup (1950, 1952), Hoar (1952) and Swift (1955, 1959) of changes in the thyroid gland, and changes induced by dosing salmonid fish with thyroid hormones, indicate that the increased activity undoubtedly shown by migrating smolts is related to the activity of this gland.

On the other hand, Hoar (1953) and Nishida (1953) relate the pre-migratory decrease in chloride to the development at this stage of numbers of ‘chloride-secretory cells’ in the gills of the two Oncorhynchus species they studied, viz. masou and nerka. This hypothesis is difficult to test. In S. salar these cells are present in the fry as soon as the gills are functional, generally some 2 years prior to migration. Whether the number of these cells increases before the seaward migration, is difficult to establish; also one might ask if the cells are non-functional during 2 years of juvenile fresh-water life. Alternatively, a suppression of the freshwater salt absorbing mechanism could occur as a pre-adaptation to marine osmoregulation, especially as certain other biochemical characteristics of the young fish change at this time to those characteristic of the adult marine fish. It seems clear from experiments described in the literature that an adjustive period is necessary after the fish has entered sea water before it can deal adequately with the different osmotic situation. Black (1957) and Houston (1959) found independently that for chum salmon fry 36 hr. is necessary for adequate regulation to be developed in sea water; Keys (1933) found a period of about 50 hr. for Anguilla vulgaris-, Houston (1959) and I (Parry, 1958) find about 50 hr. necessary for S. gairdnerii in 50% sea water; and my figures for S. trutta of smolt size and age indicate that about 200 hr. is necessary in (50% sea water), although migrating smolts of S. salar were much better able to adjust to the salinity change. Following this adjustive phase, a second or regulative phase takes its place and regulation of the plasma and muscle concentrations brings the levels of individual ions to those characteristic of marine teleosts. Histological changes in the acidophil cells in the gills (Vickers, 1958) after transfer of fish to saline media, indicate a similar time for an adjustive, followed by a regulative phase. A more recent electron microscope study of these cells (Houston & Threadgold, 1961) could support this conclusion.

Thus, far from a condition of pre-adaptation to a marine life, which the falling chloraemia of the migrating smolt has been said to induce, there are indications that the young fish are unprepared, or very inadequately prepared, physiologically for the change, and require a considerable time before the regulative phase can be brought into play. Salmonids (and other euryhaline diadromous fish) differ from other teleosts in their ability to reach this regulative phase before they suffer ill-effects from the imbalance of ions in plasma and muscle in the adjustive phase, and to be able to regulate after the fashion of both marine and freshwater teleosts, according to the environment in which they find themselves.

The material used for the physiological studies reported here has been obtained principally from natural sources from many different parts of Britain. Without assistance from the various River Boards and other authorities this would have been impossible, and I would like to thank all those who have been concerned. In particular, I am most grateful to Messrs G. Common, Northumberland and Tyneside River Board; H. Evans, Dee and Clwyd River Board; H. Gavin, Aberdeen Harbour Board; B. C. Lincoln, Scottish Hydro Electric Board, Invergarry; and L. Stewart, Lancashire River Board.